© Neuronpedia 2026
    Privacy & TermsBlogGitHubSlackTwitterContact
    Neuronpedia logo - a computer chip with a rounded viewfinder border around it

    Neuronpedia

    APIAssistant AxisNEWCircuit TracerNEWSteerSAE EvalsExports Community BlogPrivacy & TermsContact
    1. Home
    3. Joseph Bloom · Feature Splitting for GPT2-Small
    5. GPT2-Small
    7. Residual Feature Splitting 12288
    9. 8-RES_FS12288-JB
    11. 1323
    Prev
    Next
    INDEX
    Explanations

    texts related to spiritual or philosophical reflection and introspection

    oai_token-act-pair · gpt-3.5-turboTriggered by @bot
    New Auto-Interp
    Top Features by Cosine Similarity
    Comparing With GPT2-SMALL @ 8-res_fs12288-jb
    Configuration
    jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288
    Prompts (Dashboard)
    24,576 prompts, 128 tokens each
    Dataset (Dashboard)
    Skylion007/openwebtext
    Features
    12,288
    Data Type
    torch.float32
    Hook Point
    blocks.8.hook_resid_pre
    Architecture
    standard
    Context Size
    128
    Dataset
    Skylion007/openwebtext
    Hook Point Layer
    8
    Activation Function
    relu
    Embeds
    IFrame
    Link
    Not in Any Lists

    No Comments

    Neuron Alignment
    Index
    Value
    % of L₁
    87
    +0.28
    1.4%
    756
    +0.15
    0.8%
    762
    +0.12
    0.6%
    Correlated Neurons
    Index
    P. Corr.
    Cos Sim.
    87
    +0.28
    -0.04
    756
    +0.15
    -0.02
    762
    +0.12
    0.13
    Negative Logits
    senal
    -0.78
    Ops
    -0.66
    budget
    -0.65
     Reported
    -0.65
    boost
    -0.62
     ASAP
    -0.62
    umably
    -0.61
    untled
    -0.61
    busters
    -0.60
     CTR
    -0.60
    POSITIVE LOGITS
     sorrow
    1.07
     longing
    1.04
     loneliness
    1.03
     eternity
    1.03
     eternal
    1.01
     compassion
    1.01
     courage
    0.99
     remembrance
    0.97
     joyful
    0.96
     consolation
    0.96
    Activations Density 2.173%

    No Known Activations